Many DNA-templated processes, such as replication, transcription, and DNA repair, are regulated in the context of chromatin structure, which is classified into euchromatin and heterochromatin. Euchromatin is less condensed and more easily accessed by RNA polymerase II and transcription factors, which enables active transcription. In contrast, heterochromatin maintains highly condensed chromatin regions throughout the cell cycle, impeding the access of various transcription factors and causing gene silencing [1, 2]. In higher eukaryotes, heterochromatin regions are characterized by specific histone modifications, namely the methylation of histone H3K9 and histone H3K27 [1, 3]. H3K9-methylated chromatins are bound by heterochromatin protein 1 (HP1), and this process leads to heterochromatin formation [1, 2, 4, 5].

Saccharomyces cerevisiae and Schizosaccharomyces pombe are well-studied model systems for the investigation of heterochromatin. However, between the two model species, there are many differences in the mechanism of heterochromatin formation and gene silencing. Notably, although the methylation of histone H3K9 and HP1 are well-conserved in S. pombe, neither of this histone modification and heterochromatin factor exist in S. cerevisiae [1, 2, 6]. Instead, in S. cerevisiae, the formation and maintenance of heterochromatins are regulated by the silent information regulator (SIR) complex-silencing system [2, 6, 7].

Although the mechanisms for silent chromatin formation are different in both model systems, they share several essential common features. Heterochromatin regions and heterochromatin structural proteins, the Sir2/3/4 complex in S. cerevisiae and Swi6 in S. pombe, are sequestered in several foci at the nuclear periphery [7,8,9,10,11]. After a brief introduction on the mechanism of heterochromatin formation, we will introduce the regional sequestration of both heterochromatin structural proteins and heterochromatin loci at the nuclear periphery [7,8,9,10,11]. We will further discuss how the sequestered nuclear subcompartments contribute to heterochromatin structure and gene silencing. In sequestered nuclear subcompartments, heterochromatin structural proteins and heterochromatin loci form and maintain heterochromatin structure by the following strategies: (1) heterochromatin structural proteins are oligomerized [7, 12,13,14]; (2) physical interaction between heterochromatin structural proteins and nucleosomes induces conformational changes of each other [7, 12,13,14]; (3) although not identified in S. cerevisiae, H3K9-methylated nucleosomes and Swi6 in S. pombe form phase-separated liquid condensates, which maintain distinct biochemical conditions distinguished from the outer environments [7]. Through these strategies, heterochromatin structural proteins are tightly bound to nucleosomes [13, 15]. In addition, neighboring nucleosomes are tightly linked, which enables more compacted chromatin structures and gene silencing [7, 15, 16]. Through this review, we propose the importance of regional sequestration of heterochromatin structural proteins for the formation and maintenance of heterochromatin structure by exemplifying two distinct models, S. cerevisiae and S. pombe.

Heterochromatin structural proteins are recruited to heterochromatin loci via different mechanisms in S. cerevisiae and S. pombe

In S. cerevisiae, epigenetic silencing markers conserved in metazoans have not been identified. Specifically, there are no orthologous proteins of chromatin modifiers for silencing mechanisms, such as DNA cytosine methylation, H3K9 methylation, and H3K27 methylation, in S. cerevisiae. Instead, the silent chromatin structure in S. cerevisiae is maintained through the recruitment and spreading of the SIR complex, which is composed of Sir2, Sir3, and Sir4 [6].

In S. pombe, H3K9 methylation and Swi6 lead to heterochromatin formation. H3K9 methylation is a conserved histone modification, responsible for heterochromatin formation in a multitude of organisms from fission yeast to humans [3]. Clr4 is the only H3K9 methyltransferase in S. pombe, but several H3K9 methyltransferases have been identified in mammalian cells, including SUV39H1/KMT1A, SUV39H2/KMT1B, SETDB1/KMT1E, dimeric `G9a/KMT1C-GLP (G9a-like protein)/KMT1D, and the PRDM family [17]. Drosophila melanogaster retains Su(var)3–9, G9a and SETDB1 as the H3K9 methyltransferases [18]. In Arabidopsis thaliana, KYP, SUVH5, and SUVH6 are identified as H3K9 methyltransferases [19]. In Neurospora crassa, a filamentous fungus, the H3K9 methyltransferase is Dim5 [20].

In S. cerevisiae, the SIR complex is recruited to silent loci by interaction with repressor proteins bound to heterochromatin-specific DNA elements, including silencers in silent mating loci and multiple Rap1 binding sites in telomeres [6]. In S. pombe, Swi6 is recruited to silent loci by binding to H3K9-methylated nucleosomes [1, 4, 21]. Upon recruitment, both heterochromatin structural proteins are spread from nucleation sites through self-assembly [1, 6, 22].

Recruitment of SIR proteins in silent chromatin regions is regulated by DNA elements and cognate binding proteins

In S. cerevisiae, homothallic mating (HM) loci, telomeres, and rDNA loci are the silent chromatin regions [6] (Fig. 1). The mating type of budding yeast is determined by a gene positioned at the MAT locus—either “a” or “α” (alpha) [6] (Fig. 1A). A budding yeast cell, regardless of its mating type, contains all genes for both mating types in two HM loci; these HM loci—HMRa and HMLα—are located at either side of the MAT locus and maintained in a silent state, called HM silencing (Fig. 1A) [6]. Telomeres are composed of TG1-3 repeats of 300–350 bps in length and cognate binding proteins within the chromosomal ends [6] (Fig. 1B). rDNA loci are composed of approximately 100–200 rDNA repeats (Fig. 1C). Each of the repeat includes genes coding 35S pre-rRNA transcribed by RNA polymerase I and genes coding 5S rRNA transcribed by RNA polymerase III; these regions are separated by intergenic spacer (IGS) regions containing IGS1 and IGS2 (Fig. 1C) [23, 24].

Fig. 1

Heterochromatin formation in Saccharomyces cerevisiae. A Heterochromatin formation in silent mating-type loci of S. cerevisiae. At budding yeast chromosome III, two silent mating-type loci—HMLα and HMRa—surround the mating-type (MAT) locus. Each homothallic mating (HM) locus is surrounded by two proto-silencers, E and I, which nucleate the heterochromatin assembly. Silencer elements are bound by Orc1, Rap1, and Abf1. Orc1 interacts with Sir1 and Abf1 interacts with Sir3. Through the self-reinforcing mechanism of the SIR complex composed of Sir2, Sir3, and Sir4, silent chromatin is formed at HM loci. B Heterochromatin formation in telomeres of S. cerevisiae. Telomeres consist of TG1-3 repeat regions and chromosomal ends. Chromosomal ends are bound by yKu70/80 heterodimeric complexes. Telomeric repeats contain multiple Rap1 binding sites and the SIR complex is recruited to telomeric repeats through Rap1. Rif1 competes with Sir4 for binding to Rap1. The yKu complex regulates this competition process for Sir4 recruitment and SIR complex assembly. C Heterochromatin formation in rDNA repeats of S. cerevisiae. Approximately 100 to 200 rDNA repeats are positioned at chromosome XII. Each repeat consists of the 35S pre-rRNA gene and 5S rRNA gene which are separated by intergenic spacer 1(IGS1). IGS2 is located upstream of the 5S rRNA gene. A replication fork barrier (RFB) site is positioned within IGS1 and the binding site for Fob1. The binding of Fob1 into RFB sites causes recombination of rDNA repeats, which should be prevented by the binding of additional proteins. Net1 tethers to Fob1 and recruits Cdc14 and Sir2 into rDNA loci, thereby forming the regulator of nucleolar silencing and telophase exit (RENT) complex. Tof2 binds to Fob1, leading to the recruitment of two cohibin complex components, Lrs4 and Csm1, for Sir2-independent rDNA silencing. Lrs4/Csm1 interacts with Heh1/Nur1 (two nuclear membrane proteins)

Specific DNA elements in HM loci, telomeres, and rDNA loci are bound by cognate binding proteins, which recruit SIR proteins to silent chromatin regions. Silent HM loci are positioned between two silencers, E and I, and silencers are bound by repressor and activator protein 1 (Rap1), autonomously replicating sequences (ARS) binding factor 1 (Abf1), and origin replication complex (ORC) [25, 26] (Fig. 1A). These silencer-binding proteins form a loop structure and recruit the SIR complex to nucleate silent chromatin formation [26,27,28]. For example, Abf1 binds to Sir3, and ORC interacts with Sir1, thereby bringing Sir4 into silent loci [6, 29, 30]. After being recruited to chromatin via interaction with Sir4, Sir2, which is a protein of the NAD+-dependent histone deacetylase (HDAC) family, deacetylates acetylated H4K16 (H4K16ac) [31,32,33,34]. Deacetylation of acetylated H4K16 enhances the access of Sir3 to chromatin and blocks H3K79 methylation [31,32,33,34]. In telomeres, Rap1 and the yKu70/80 heterodimeric complex are bound to multiple Rap1 binding sites in TG1-3 repeats and chromosomal ends, respectively [35]. The Sir3–Sir4 dimer is recruited to subtelomere regions by binding to the carboxy-terminal domain of Rap1 [36, 37]. The interaction between Rap1 and Sir4 is inhibited by Rif1, and the yKu70/80 heterodimeric complex contributes to the binding of Sir4 to Rap1 by suppressing the inhibitory effect of Rif1 [35]. In rDNA loci, the mechanism of silent chromatin formation at the IGS1 region has been studied more than that at IGS2 (Fig. 1C). The replication fork barrier (RFB) site is positioned within IGS1 and bound by Fob1 [23]. Net1 is bound to Fob1 and brings Cdc14 and Sir2 to form the regulator of nucleolar silencing and telophase exit (RENT) complex [38, 39]. Topoisomerase associated factor 2 (Tof2) is bound to Fob1 and brings two cohibin complex components (Lrs4 and Csm1) [23, 40, 41]. Lrs4 and Csm1 directly recruit the condensin complex and enable correct alignment between sister chromatids [42, 43]. Therefore, unequal sister chromatid exchange is prevented, and the stability of rDNA regions is increased by the condensin complex [42].

In S. pombe, H3K9 methylation domains are formed at heterochromatin regions to initiate Swi6 recruitment

H3K9 methylation domains in heterochromatin regions are important for the chromatin recruitment of Swi6 and are initiated in heterochromatin nucleation sites [4, 21]. The representative heterochromatic regions in S. pombe are pericentromeric regions as well as telomeres and silent mating-type loci (Fig. 2). S. pombe possesses 10-kb centromeric regions, which are large compared to the much smaller 125-bps centromeres in S. cerevisiae (Fig. 2A) [44]. The central region of the centromeres is composed of two classes of DNA sequences, imr (innermost repeat) and cnt (central) repeats. The cnt region is flanked by two imr sequences in an inverted orientation [1, 45]. This central region, consisting of cnt and two imr sequences, is flanked by otr (outer repeat) repeats [46]. This pericentromeric otr repeats contain dg and dh repeat sequences, which are the RNAi-dependent heterochromatin initiation sites [46]. The mating type of fission yeast, either + or −, is determined according to the gene—either mat2P or mat3M—located at the MAT locus. Like S. cerevisiae, both mat2P and mat3M elements are maintained in a transcriptionally silent state (Fig. 2B) [47, 48]. Two heterochromatin initiation sites, CenH region and REIII elements, are located between mat2P and mat3M. The CenH region is composed of multiple repeats homologous to centromeric dg/dh repeats [45]. In telomeres, chromosomal ends contain approximately 300 bps of telomeric double-stranded DNA repeats, with a single-stranded overhang that protrudes from the ends (Fig. 2C) [49]. Telomere-associated sequences (TAS) refer to the DNA regions proximal to telomeric DNA repeats [49,50,51]. In addition, multiple CenH-like regions were identified in the more distal regions to telomeric DNA repeats [45, 50].

Fig. 2

Heterochromatin formation in Schizosaccharomyces pombe. A Heterochromatin formation in centromeres. In centromeres, centromeric repeats (cnt) are surrounded by two innermost repeats (imr) regions with inverted orientation. Outer repeat (otr) elements are positioned further outside, and they consist of dg/dh repeat elements. dg/dh repeats nucleate heterochromatin in an RNAi-dependent manner. Once nucleated, heterochromatin is spread through the H3K9me-Swi6-dependent self-reinforcing mechanism up to its encounter with boundary elements including tRNA gene clusters. B Heterochromatin formation in silent mating-type loci. The mating-type is determined by the MAT1 gene; two mating-type determining regions, mat2P and mat3M, which contain information on both mating types, are maintained in silent chromatin. CenH regions show high homology to centromeric dg/dh repeats and function as RNAi-dependent nucleation centers. The REIII element is the binding site for Atf1/Pcr1, constituting an RNAi-independent heterochromatin nucleation site. Once nucleated, heterochromatin is spread up to its encounter with boundary elements, IR (inverted repeats)-L and IR-R. C Heterochromatin formation in telomeres. Telomeres are composed of double-stranded telomeric DNA repeats and single-stranded overhangs at chromosomal ends. Telomeric DNA repeats and immediately adjacent telomere-associated sequences (TAS) are RNAi-independent heterochromatin nucleation sites. Regions more distal to chromosomal ends contain multiple CenH-like sequences, functioning as RNAi-dependent heterochromatin nucleation sites (e.g., centromeres and silent mating-type loci). The shelterin complex consists of Tpz1/Pot1 subcomplexes bound to single-stranded overhangs and the double-stranded telomeric repeat-binding protein Taz1 connected by Rap1 and Poz1. The shelterin component Ccq1 recruits Clr4 for H3K9 methylation and subsequent Swi6-dependent heterochromatin formation. Additionally, Ccq1 leads to SHREC recruitment for transcriptional gene silencing. There is no known boundary element at telomeres; therefore, telomeres contain a long transition zone showing a gradual decrease in the heterochromatin domain

In centromeres, the recruitment of Clr4 and histone deacetylases (HDACs) into centromeric dg and dh repeats to initiate heterochromatin formation is mediated by RNAi-dependent and -independent mechanisms [52,53,54,55] (Fig. 2A). In silent mating loci, RNAi-dependent heterochromatin assembly starts at the CenH region, which is homologous to centromeric dg/dh repeats, and RNAi-independent heterochromatin assembly is initiated at REIII elements by the cognate binding proteins Atf1/Pcr1 [1, 56]. Similar to the Atf1/Pcr1-dependent heterochromatin assembly in silent mating-type loci, telomeres form and maintain heterochromatin through the shelterin complex (Fig. 2C) [49, 51, 53,